基于方向区域的NSCT图像融合算法

提出一种新的基于方向区域的NSCT图像融合算法。算法首先对源图像进行NSCT分解,获得不同方向的高低频子带。其次对高低频系数,根据不同分解层的方向特性,按方向区域能量的规则进行融合。最后,通过反变换获得融合图像。该方法既保留了Contourlet变换方法的多尺度多方向特性,又具有移不变性。实验结果表明,提出的算法有效可行,对比常用的区域融合算法,获得了更好的融合效果。

Nonsubsampled contourlet image fusion algorithm based on directional region

A novel nonsubsampled contourlet image fusion algorithm based on directional region is presented.Firstly,the source images are decomposed into multidirectional highpass and lowpass subbands by nonsubsampled contourlet transform.Secondly,fusion coefficients are selected according to the directional region energy.Finally,the fused image is obtained through inverse transformation.Similar to the contourlet transform method,the result of this approach is multiscalar and multidirectional image decomposition with good multiresolution.It is also shift-invariant.Experimental results show that the proposed method outperforms the traditional transform and improves the quality of the fused image.